CN117114891A

CN117114891A - Transaction strategy income prediction method, system, electronic equipment and storage medium

Info

Publication number: CN117114891A
Application number: CN202311114369.0A
Authority: CN
Inventors: 徐祎; 金戈
Original assignee: Alipay Hangzhou Information Technology Co Ltd
Current assignee: Alipay Hangzhou Information Technology Co Ltd
Priority date: 2023-08-31
Filing date: 2023-08-31
Publication date: 2023-11-24

Abstract

The embodiment of the specification discloses a transaction strategy income prediction method, a transaction strategy income prediction system, electronic equipment and a storage medium. In the trading strategy profit prediction method, based on return measurement prediction profit corresponding to each market data prediction model and return measurement real profit corresponding to financial trading strategy, a preferred market data prediction model matched with the financial trading strategy is selected instead of a market data prediction model with highest comprehensive prediction accuracy of the market data, market prediction data in a future preset time period are obtained according to the preferred market data prediction model in a prediction mode, and future prediction profit using the financial trading strategy in the future preset time period is obtained according to the future preset time period market prediction data and the financial trading strategy.

Description

Transaction strategy income prediction method, system, electronic equipment and storage medium

Technical Field

One or more embodiments of the present disclosure relate to the field of mobility management technology, and in particular, to a transaction policy profit prediction method, system, electronic device, and storage medium.

Background

In the trade process of the financial homonymy, mobility management of the assets is particularly important, a part of the assets can flow in and out as the fund-financing property assets to generate benefits, a part of the assets can be bought into or sold out of commodities in the financial market as the financial property assets, such as bonds, securities, futures and the like, and the benefits obtained by the financial property assets can be further used as mobile funds to further generate benefits.

Therefore, if asset value-added is to be achieved in the mobility management of assets, and the mobility stability of the assets is guaranteed, the transaction policy formulation for the assets with financial properties is particularly important, and the key of the transaction policy formulation is the yield prediction of the transaction policy.

Disclosure of Invention

The embodiment of the specification provides a transaction policy income prediction method, a system, electronic equipment and a storage medium, wherein the technical scheme is as follows:

in a first aspect, embodiments of the present disclosure provide a transaction policy benefit prediction method, including:

predicting a plurality of different return time period market prediction data in the same return time period by using a plurality of market data prediction models;

according to the market history data of the return time period and the financial transaction strategy, the return real benefits corresponding to the financial transaction strategy are obtained;

According to the market forecast data and financial transaction strategies of a plurality of different return time periods, return forecast benefits corresponding to the market data forecast models are obtained;

selecting a preferable market data prediction model matched with the financial transaction strategy according to each return predicted return and return real return;

predicting and obtaining market prediction data of a future preset time period according to a preferable market data prediction model;

and according to the market forecast data of the future preset time period and the financial transaction strategy, obtaining future forecast benefits of using the financial transaction strategy in the future preset time period.

In a second aspect, embodiments of the present disclosure provide a trading strategy revenue prediction system, comprising:

the market data prediction module is used for predicting and obtaining a plurality of different return time period market prediction data in the same return time period by using a plurality of market data prediction models;

the profit calculation module is used for obtaining the return real profit corresponding to the financial transaction strategy according to the return time period market history data and the financial transaction strategy; the system is also used for obtaining the return prediction benefits corresponding to the market data prediction models according to the market prediction data and financial transaction strategies of a plurality of different return time periods;

The model matching module is used for selecting a preferable market data prediction model matched with the financial transaction strategy according to each return predicted return and return real return;

the market data prediction module is further used for predicting and obtaining market prediction data of a future preset time period according to the optimal market data prediction model;

the profit calculation module is further used for obtaining future predicted profits by using the financial transaction strategy in a future preset time period according to the future preset time period market prediction data and the financial transaction strategy.

In a third aspect, embodiments of the present disclosure provide an electronic device including a processor and a memory; the processor is connected with the memory; the memory is used for storing executable program codes; the processor executes a program corresponding to the executable program code by reading the executable program code stored in the memory, for executing the steps of the transaction policy benefit prediction method according to the first aspect of the above embodiment.

In a fourth aspect, embodiments of the present disclosure provide a computer storage medium storing a plurality of instructions adapted to be loaded by a processor and to perform the steps of the transaction policy benefit prediction method of the first aspect of the embodiments described above.

The technical scheme provided by some embodiments of the present specification has the following beneficial effects:

by means of the back measurement, based on the back measurement predicted return corresponding to each market data prediction model and the back measurement real return corresponding to the financial transaction strategy, a preferable market data prediction model matched with the financial transaction strategy is selected instead of the market data prediction model with highest comprehensive market data prediction accuracy, so that the transaction strategy return rate prediction accuracy of the financial property asset is improved, a foundation is laid for establishing the transaction strategy of the financial property asset, and better asset liquidity management is achieved.

Drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present description, the drawings that are required in the embodiments will be briefly described below, and it is apparent that the drawings in the following description are only some embodiments of the present description, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.

Fig. 1 is a schematic diagram of an application scenario of a trading strategy profit prediction system according to an embodiment of the present disclosure.

Fig. 2 is a flow chart of a transaction policy benefit prediction method according to an embodiment of the present disclosure.

FIG. 3 is a flow chart of yet another method for predicting trading strategy revenue according to an embodiment of the present disclosure.

Fig. 4 is a flow chart of another method for predicting trading strategy revenue according to an embodiment of the present disclosure.

Fig. 5 is a schematic structural diagram of a trading strategy profit prediction system according to an embodiment of the present disclosure.

FIG. 6 is a schematic diagram of a further trading strategy revenue prediction system provided by embodiments of the present disclosure.

FIG. 7 is a schematic diagram of a transaction policy benefit prediction system according to one embodiment of the present disclosure.

Fig. 8 is a schematic structural diagram of an electronic device according to an embodiment of the present disclosure.

Detailed Description

The technical solutions in the embodiments of the present specification will be clearly and completely described below with reference to the drawings in the embodiments of the present specification.

The terms first, second and the like in the description and in the claims and in the above-described figures are used for distinguishing between different objects and not necessarily for describing a sequential or chronological order. Furthermore, the term "include" and any variations thereof is intended to cover a non-exclusive inclusion. For example, a process, method, system, article, or apparatus that comprises a list of steps or elements is not limited to only those listed steps or elements but may include other steps or elements not listed or inherent to such process, method, article, or apparatus.

The present specification, prior to detailing a transaction policy benefit prediction method in connection with one or more embodiments, describes a scenario in which the transaction policy benefit prediction method is applied.

The transaction policy and income prediction method provided by the embodiments of the present disclosure can be applied to a scenario of financial peer transaction, where financial peer transaction refers to the fund flow and transaction activities between financial institutions. These financial institutions may include commercial banks, securities companies, insurance companies, investment funds, etc., between which funds flow and trade activities are conducted through various channels.

Financial trade in the trade mainly includes the following two aspects:

and (5) fund melting: and the financial homonymy obtains the income by providing borrowing and loaning for the opposite party.

Financial transactions: financial peers may conduct trading activities on securities markets or bond markets, including buying and selling stocks, bonds, futures, etc., to obtain revenue.

It should be noted that, the financial trade in the same industry may also include other trade types, which can be included in both aspects of money-melting and financial trade, for example, foreign exchange trade can be included in financial trade, and will not be described in detail herein.

Liquidity management refers to a series of policies and measures taken by a financial institution or investor to ensure availability and flexibility of their funds to meet various funds needs and to cope with liquidity risks of the market to ensure that the financial institution or investor is able to meet cash inflow and outflow needs.

It will be appreciated that during a financial homography transaction, a portion of the assets will be used as funds-melting property assets to generate revenue and a portion will be used as financial property assets to generate revenue, and it will be appreciated that the revenue obtained by the financial property assets can be further used as flowing funds to further generate revenue. Therefore, if asset value-added is to be achieved in the mobility management of assets, and the mobility stability of the assets is guaranteed, the transaction strategy formulation aiming at the assets with financial properties is particularly important, and the key of the transaction strategy formulation is the yield prediction of the transaction strategy. Therefore, the trading strategy income prediction method provided by the embodiments of the present disclosure can improve the trading strategy income prediction accuracy of the financial attribute asset, lay a foundation for the trading strategy formulation of the financial attribute asset, and further realize better asset liquidity management.

It can be appreciated that the transaction policy and income prediction method provided in the embodiments of the present disclosure may be used by not only each financial institution in a financial trade scenario, but also by scattered households. The dispersed user also has an asset account, a part of assets in the asset account are used for various expenses, and meanwhile, the dispersed user also has an asset account entry, so that the part of assets can be regarded as the above-mentioned fund-fusion attribute assets, or the part of assets of the dispersed user and the fund-fusion attribute assets of the financial institution can be collectively called as non-financial asset, only the part of assets of the dispersed user do not generate extra income, and the rest of assets in the dispersed user asset account are usually used for financial management, so that the transaction strategy income prediction method provided by the embodiment of the present disclosure can also provide the dispersed user for use.

Next, referring to fig. 1, a schematic application scenario of a transaction policy profit prediction system provided by the embodiment of the present disclosure is shown, and the transaction policy profit prediction system provided by the embodiment of the present disclosure is explained by a mobility management scenario in a financial peer transaction.

As shown in fig. 1, the application scenario of the transaction policy profit prediction system may at least include an electronic device 101, where:

the electronic device 101 is provided with a third party application program capable of conducting transaction management, where the third party application program may be, but not limited to, an application program such as a browser, and a specific example is a certain application program:

the administrator may click on the application on the display interface of the electronic device 101 and pass authentication to log in to the application. It should be noted that the application should be bound to a financial institution's funds account, so that the application may obtain various funds conditions in the financial institution's funds account, especially the funds-melting property asset flow history and asset balance conditions. The application may also obtain market data and simulate the revenue of the trading strategy under the corresponding market data. The history of movement of the property of the financing property is the history of movement of borrowing, loan and the like. Market data includes the price of stock, bond, futures, etc.

After logging in the application program, the administrator can know each fund condition of the financial institution in the application program, and can input a transaction strategy needing to predict future benefits on an application program interface and click a future prediction benefit generation button, at this time, the application program background can generate the future prediction benefits of the transaction strategy through the transaction strategy benefit prediction method described in one or more embodiments below, and the future prediction benefits of the transaction strategy are displayed on the application program interface so as to provide references for the administrator to formulate the future transaction strategy.

It will be appreciated that if there are a plurality of transaction policies for predicting future benefits, after the administrator clicks the future prediction benefit generation button, a reference table or picture with each transaction policy specific content and the future prediction benefits corresponding to each transaction policy is displayed on the application program interface. And the transaction strategy with the highest future predicted gain in various transaction strategies can be marked so that an administrator can quickly find the preferred financial transaction strategy.

It should be noted that the transaction policy may also be automatically generated in the background of the application program, without being input by an administrator.

It will be appreciated that the scenario of the trade strategy benefit prediction by the trade strategy benefit prediction system is similar to that described above, and will not be described in detail herein.

The electronic device 101 referred to in the embodiments of the present description may be a smart phone, tablet, desktop, laptop, notebook, ultra-mobile personal computer (UMPC), handheld computer, PC device, personal digital assistant (Personal Digital Assistant, PDA), virtual reality device, etc.

The transaction policy benefit prediction methods provided in the embodiments of the present disclosure all relate to a time sequence loop method, so before the transaction policy benefit prediction method is elaborated in connection with one or more embodiments, the time sequence loop method related to the transaction policy benefit prediction method is introduced:

timing loop (Time Series Backtesting) is a method for evaluating investment strategies or trading algorithms. It evaluates past performance of policies by modeling transactions on historical data and infers their potential effect in the future.

Referring to fig. 2, fig. 2 is a flow chart illustrating a transaction policy benefit prediction method according to an embodiment of the present disclosure, where the method may be performed by the electronic device shown in fig. 1.

As shown in fig. 2, the transaction policy profit prediction method may at least include the following steps:

step 202, predicting and obtaining a plurality of different time period market prediction data in the same time period by using a plurality of market data prediction models;

step 204, according to the market history data of the return time period and the financial transaction strategy, obtaining return real benefits corresponding to the financial transaction strategy;

step 206, obtaining return prediction benefits corresponding to the market data prediction models according to a plurality of different return time period market prediction data and financial transaction strategies;

step 208, selecting a preferable market data prediction model matched with the financial transaction strategy according to each return predicted benefit and return real benefit;

step 210, predicting and obtaining market prediction data of a future preset time period according to a preferable market data prediction model;

step 212, obtaining future prediction benefits of using the financial transaction strategy in the future preset time period according to the market prediction data and the financial transaction strategy in the future preset time period.

It will be appreciated that the accuracy of future revenue calculation for a certain trading strategy is closely related to the accuracy of future market data predictions, but the higher the overall accuracy of future market data predictions, the more accurate the revenue prediction for a certain trading strategy is, because there is a temporal concern for distinction from a statistical value concern, for example:

For example, a certain transaction policy is: if the stock price is more than 10 yuan at 11 am every day, 1000 shares of the stock are sold, and if the stock price is less than 9 yuan at 4 pm every day, 1000 shares of the stock are purchased. It will be appreciated that the time nodes of interest in the transaction policy are 11 am and 4 pm, respectively. It will thus be appreciated that the market data prediction model employed may be inaccurate for other time nodes of the market data prediction model, but the prediction data at 11 am and 4 pm need to be accurate for predicting future benefits of the trading strategy.

For another example, a transaction policy is: if the maximum value of the stock price exceeds 10 yuan in the future preset time length, the 1000 stocks of the stock are sold, and if the minimum value of the stock price is lower than 8 yuan in the future preset time length, the 1000 stocks of the stock are purchased.

It will be appreciated that the statistics of interest in the transaction strategy are maximum and minimum, respectively. It will thus be appreciated that other statistics of the market data prediction model employed may be inaccurate, such as average stock price, but the maximum and minimum possible stock prices need to be predicted accurately, in order to predict future benefits of the trading strategy.

That is, the market data prediction model adopted needs to accurately predict the data concerned by the trading strategy, so that the income of the trading strategy can be accurately predicted.

However, it will be appreciated that the transaction policy for a period of time typically includes a plurality of sub-policies, and each sub-policy may be set in sequence, for example, 1000 strands may be sold if the maximum value of the stock price is greater than a preset value from 9 am to 11 am, and 1000 strands may be sold if the average value of the stock price is greater than a preset value from 2 pm to 5 pm. Overlapping settings are also possible, such as selling 1000 strands if there is a maximum value of stock price greater than a preset value at 9 am to 11 am, or selling 1000 strands if the average value of stock price is greater than a preset value at 9 am to 11 am. In this way, it is difficult to evaluate in advance which prediction model is more suitable for the transaction policy, and it may be that the prediction model obtained through training all the statistical features is more accurate, or that the prediction model obtained through training only the maximum statistical features is more accurate, so that the prediction models suitable for the transaction policy are not screened in advance in the embodiments of the present specification, but prediction benefits corresponding to each prediction model are obtained through calculation by traversing each prediction model, so that the screening of the preferred market data prediction model is performed.

In various embodiments of the present disclosure, there is at least one of an internal architecture setting, a statistical feature extracted during training, a learning algorithm employed during training, and training data employed for training between any two market data prediction models of the plurality of market data prediction models.

It can be understood that only the internal architecture setting may be different between any two market data prediction models, or the internal architecture setting may be different, and the statistical features extracted in the training process may be different, or the four items may be different.

It will be appreciated that if the same data is input to different models, there will be a difference in the output of each model, since different models may have different architectures, and the statistical features extracted during training, the learning algorithm employed, and the training data may also be different, so that when these predictive models are used, they will also have differences in the processing mode and capturing capability of the data, and therefore the final output data will also be different.

The following description will be made by taking different statistical characteristics extracted in the training process as an example:

if the training targets are: training to obtain a prediction model which can better predict the maximum strand value possibly occurring in the preset time in the future, and adding the maximum statistical feature into the model training process.

If the training targets are: training to obtain a prediction model which can better predict the average value of stock price in the preset time in the future, and adding the average value statistical characteristics into the model training process.

In one embodiment of the present description, the statistical features extracted during the training of the market data predictive model include one or more of maximum, minimum, mean, standard deviation. One of the extraction may be performed, or a plurality of extraction may be performed simultaneously.

The following description will be made by taking different training data adopted in the training process as an example:

it will be appreciated that the predictive power of the model will be different for different training data, e.g. more data for 9 am in the training data, and thus the accuracy of the trained model will be higher for 9 am.

The following description will take the difference of the internal architecture settings of the model as an example:

the market data prediction model described in the various embodiments of the present disclosure may employ existing four time series prediction models, such as a tree model, a linear regression model, a bayesian structured time series model, and a transform-based deep learning model. The following is a brief description of the data processing advantages of each of the four time sequence prediction models:

Tree model: the tree model is suitable for processing classification problems and regression problems, and has certain robustness for missing values and outliers of the features.

Linear regression model: the linear regression model is suitable for dealing with regression problems of continuous target variables.

Bayesian structured timing model: the method can capture dynamic relations in time series data, can carry out probability inference and prediction, can consider priori knowledge and uncertainty of the data, and is beneficial to processing the situations of data missing and noise.

Deep learning model based on transducer: the transducer model is capable of capturing long-range dependencies in a sequence.

It will be appreciated that different time series prediction models have different architectures, so that there are different advantages in processing data, and further that even if the same data is input, the data output by these models will be different.

In an embodiment of the present disclosure, the learning algorithm adopted in the training process may be any one of a gradient descent algorithm, a random gradient descent algorithm, a batch gradient descent algorithm, a momentum optimization algorithm, an adaptive learning rate optimization algorithm, and an L-BFGS algorithm, and different learning algorithms are adopted to perform model training, so that the principle of outputting different models is similar to the three factors, and will not be repeated herein.

In various embodiments of the present disclosure, the selection of the adaptive prediction model for the financial transaction policy is verified by a time sequence loop method. First, market history data needs to be acquired, and the market history data is divided into training market history data and return time period market history data. And training the market data prediction model by further adopting training market history data to obtain a trained market data prediction model. Further, based on the market history data of the return time period and the financial transaction strategy for predicting future benefits, return real benefits corresponding to the financial transaction strategy are obtained. Further, the return time period market prediction data is predicted based on the post-training market data prediction model, and it can be understood that the training data when the market data prediction model is trained does not include the return time period market history data, so that a certain difference exists between the return time period market prediction data and the return time period market history data based on the post-training market data prediction model. Further, according to a plurality of different market prediction data and financial transaction strategies in the time slot for return, the return prediction benefits corresponding to the market data prediction models are obtained, and it can be understood that, because there is a certain difference between the market prediction data in the time slot for return and the market history data in the time slot for return, there is a certain difference between the return prediction benefits and the return real benefits, the return prediction benefits are higher than the return real benefits, the return prediction benefits are lower than the return real benefits, and there is a certain possibility that the return prediction benefits are consistent with the return real benefits. Finally, according to each return predicted return and return real return, a preferable market data prediction model adapted to the financial transaction strategy can be selected, and it is understood that the absolute value of the difference between the return predicted return and the return real return corresponding to the preferable market data prediction model should be the smallest.

Moreover, it should be noted that, in the embodiments of the present disclosure, the return measured real returns and the return measured predicted returns may be obtained by performing a simulation transaction on the financial transaction policy according to time based on the corresponding market data, so as to perform a simulation calculation.

It should also be noted that market data generally includes:

stock market data: including the price of stock for opening, the price of closing, the highest price, the lowest price, the volume of delivery, the market value, etc.

Bond market data: including the amount of bond issued, expiration time, ticket interest rate, yield, bond rating, etc.

Foreign exchange market data: including exchange rates for monetary pairs, purchase prices, sell bids, transaction amounts, and the like.

Futures market data: including the trading price, volume, holding capacity, date of delivery, etc. of futures contracts.

Option market data: including the purchase price, the bid price, the line price, the expiration time, the implied volatility, etc. of the option contract.

Derivative market data: including price, volume of contact, volatility, etc. of various derivative contracts.

Fund market data: including net value, share, rate of return, portfolio, etc.

And when the income is predicted according to the transaction strategy, the used market data can be subjected to data preprocessing, including preprocessing such as data cleaning, data conversion, feature selection, feature extraction, data standardization and the like.

Referring to fig. 3, fig. 3 is a schematic flow chart of a transaction policy benefit prediction method according to another embodiment of the present disclosure, in which only steps 314 to 320 described below are shown, and steps 302 to 312 described below are consistent with steps 202 to 212 described above, so that the method is not repeated in fig. 3, and may be performed by the electronic device shown in fig. 1.

As shown in fig. 3, the transaction policy profit prediction method may at least include the following steps:

step 302, predicting and obtaining a plurality of different time period market prediction data in the same time period by using a plurality of market data prediction models;

step 304, according to the market history data of the return time period and the financial transaction strategy, obtaining return real benefits corresponding to the financial transaction strategy;

step 306, obtaining return prediction benefits corresponding to the market data prediction models according to a plurality of different return time period market prediction data and financial transaction strategies;

step 308, selecting a preferable market data prediction model matched with the financial transaction strategy according to each return predicted benefit and return real benefit;

step 310, predicting and obtaining market prediction data of a future preset time period according to a preferable market data prediction model;

Step 312, obtaining future prediction benefits of using the financial transaction strategy in the future preset time period according to the future preset time period market prediction data and the financial transaction strategy;

the steps 302 to 312 can be referred to as steps 202 to 212, and are not described herein.

Step 314, selecting a preferable market data prediction model adapted to other financial transaction strategies according to the same method;

step 316, predicting and obtaining market prediction data of a preset time period in the future according to a preferable market data prediction model adapted to other financial transaction strategies;

step 318, predicting the obtained market prediction data of the future preset time period and other financial transaction strategies according to the prediction model of the preferred market data respectively adapted to the other financial transaction strategies so as to obtain future prediction benefits of using the other financial transaction strategies in the future preset time period;

step 320, selecting and obtaining a preferred financial transaction strategy used in a future preset time period according to the future prediction benefits corresponding to all the financial transaction strategies.

The financial transaction strategy profit prediction method described in the embodiments of the present disclosure is for making a financial transaction strategy with higher profit margin in a preset time period in the future, so as to perform better asset mobility management. Therefore, after step 312, the same method as the above steps 302 to 312 can be adopted to select a preferred market data prediction model adapted to other financial transaction strategies for other financial transaction strategies; predicting and obtaining market prediction data of a future preset time period according to a preferable market data prediction model which is adapted to other financial transaction strategies; according to the market forecast data of the future preset time period and other financial transaction strategies which are forecast by the forecast model of the preferable market data respectively adapted to other financial transaction strategies, the future forecast benefits of using other financial transaction strategies in the future preset time period are obtained; and then selecting and obtaining a preferred financial transaction strategy used in a future preset time period according to the future prediction benefits corresponding to all the financial transaction strategies.

It will be appreciated that the other financial transaction strategies described above may be one or more. The preferred financial transaction strategy is one of all transaction strategies with the highest future predicted return.

Referring to fig. 4, fig. 4 is a schematic flow chart of a transaction policy benefit prediction method according to still another embodiment of the present disclosure, in which only steps 422 to 424 described below are shown, and steps 402 to 420 described below are consistent with steps 302 to 320 described above, so that the method is not repeated in fig. 4, and may be performed by the electronic device shown in fig. 1.

As shown in fig. 4, the transaction policy profit prediction method may at least include the following steps:

step 402, predicting and obtaining a plurality of different time period market prediction data in the same time period by using a plurality of market data prediction models;

step 404, obtaining return real benefits corresponding to the financial transaction strategy according to the return time period market history data and the financial transaction strategy;

step 406, obtaining return prediction benefits corresponding to the market data prediction models according to a plurality of different return time period market prediction data and financial transaction strategies;

step 408, selecting a preferable market data prediction model adapted to the financial transaction strategy according to the predicted return and the real return;

Step 410, predicting and obtaining market prediction data of a future preset time period according to a preferable market data prediction model;

step 412, obtaining future prediction benefits of using the financial transaction strategy in the future preset time period according to the future preset time period market prediction data and the financial transaction strategy;

step 414, selecting a preferable market data prediction model adapted to other financial transaction strategies according to the same method;

step 416, predicting and obtaining market prediction data of a future preset time period according to a preferred market data prediction model adapted to other financial transaction strategies;

step 418, predicting the obtained market prediction data of the future preset time period and other financial transaction strategies according to the optimal market data prediction model respectively adapted to the other financial transaction strategies so as to obtain future prediction benefits of using the other financial transaction strategies in the future preset time period;

step 420, selecting and obtaining a preferred financial transaction strategy used in a future preset time period according to the future prediction benefits corresponding to all financial transaction strategies;

the steps 402 to 420 can refer to the steps 302 to 320, and are not described herein.

Step 422, predicting and obtaining non-financial asset liquidity prediction data in a future preset time period according to the non-financial asset liquidity history data;

Step 424, obtaining the predicted cost of using the preferred financial transaction strategy in the future preset time period according to the preferred financial transaction strategy and the market prediction data of the future preset time period corresponding to the preferred financial transaction strategy;

step 426, proportional adjustment is performed on the preferred financial transaction strategy according to the asset balance, the non-financial asset movement prediction data, and the predicted cost of the preferred financial transaction strategy.

Based on the foregoing, the embodiments of the present disclosure provide a transaction policy profit prediction method, which is ultimately used for better asset liquidity management, and the basic requirement of asset liquidity management is that asset liquidity is stable. From the foregoing description, it will be appreciated that non-financial asset liquidity forecast data, i.e., the financial institution's capital-liquidity property assets, or the expense or expense-in property assets of the free account, if to ensure capital-liquidity property asset liquidity stability, or to ensure expense and expense-in property liquidity stability of the free account, the asset is not excessively mobilized as the financial attribute asset but the maximum profitability is realized, so that after step 420, the method further comprises predicting and obtaining the non-financial asset liquidity prediction data in a future preset time period according to the non-financial asset liquidity history data; further obtaining the predicted cost of using the preferred financial transaction strategy in the future preset time period according to the preferred financial transaction strategy and the market prediction data of the future preset time period corresponding to the preferred financial transaction strategy; and then, proportional adjustment is carried out on the preferred financial transaction strategy according to the asset balance, the non-financial asset liquidity prediction data and the prediction cost of the preferred financial transaction strategy.

It should be noted that, the non-financial asset flow prediction data is the asset amount to be flown out or the asset amount to be flown in the future preset time period, and the asset balance includes all the assets except the assets which have been flown out as the non-financial asset, and includes the assets which have been used as the financial attribute asset, it is understood that, if the financial attribute asset is recovered after each time period, there is no asset which has been used as the financial attribute asset.

The following examples are given:

for example, in a future preset time period, the flow prediction data of the non-financial asset is 1000 ten thousand yuan, the current balance of the asset is 2000 ten thousand yuan, and the amount of the asset required to be used by the preferred financial transaction strategy is 2000 ten thousand yuan.

It will be appreciated that if performed in accordance with the presently preferred financial transaction strategy, a steady flow of funds will not be satisfied for a predetermined period of time in the future. It is therefore necessary to adjust the preferred financial transaction strategy so that a steady flow of funds is ensured over a predetermined period of time in the future. The specific adjustment method can be, but not limited to, proportional adjustment, namely, the buying and selling amounts of the preferred financial transaction strategy are proportionally adjusted (note: the rest of the preferred financial transaction strategy is not changed), so that the total proportional adjustment of the amount of assets required to be used by the preferred financial transaction strategy can be performed, for example, the buying and selling amounts in the preferred financial transaction strategy are 1000 strands and 500 strands respectively, and at the moment, the buying and selling amounts can be adjusted to 500 strands and 250 strands are sold. It will be appreciated that the use of scaling in this embodiment may preserve the profitability of the preferred financial transaction strategy prior to scaling.

In one embodiment of the present disclosure, the non-financial asset flow forecast data is a span of intervals, thus providing a better reference value for an administrator or a bulk (user) to adjust the preferred financial transaction strategy. The non-financial asset mobile prediction data is predicted by any one of a probability prediction method, a residual fitting prediction method and a quantile regression prediction method, and the probability prediction method, the residual fitting prediction method and the quantile regression prediction method are simply introduced as follows:

probability prediction method (Probabilistic Forecasting): the probability prediction method is a method of predicting by establishing a probability model, which does not give a single predicted value but describes uncertainty by establishing a probability distribution. Common probabilistic predictive methods include bayesian statistical models, time series models such as ARIMA and GARCH, and machine learning methods such as random forests and neural networks. The probabilistic predictive method may provide more comprehensive predictive results, including point estimates and confidence intervals, etc., and thus may enable non-financial asset flow prediction data to be output as a range of intervals.

Residual fitting prediction method (Residual Fitting and Prediction): the residual fitting prediction method is a fitting and prediction method based on historical data. It first predicts the target variable with a model and then calculates the residual (difference) between the actual observed value and the model predicted value. The residuals are used to adjust the model predicted value so that the target variable is output as a range of intervals, and the prediction accuracy is improved. This approach can help correct the model bias and improve the prediction accuracy.

Quantile regression prediction method (Quantile Regression Forecasting): the quantile regression prediction method is a prediction method based on a quantile regression model. Unlike conventional regression models which only focus on predicting the mean value of the target variable, quantile regression models can predict the value of the target variable at different quantiles. By estimating a plurality of quantiles, more detailed prediction results including a median, upper and lower quantiles, etc. can be obtained, and thus the non-financial asset movement prediction data can be outputted as a range of intervals. The method has the advantages in the aspects of potential asymmetric distribution, tail risk, abnormal value and the like, is commonly used in the fields of financial risk management, market prediction and the like,

it can be understood that, different prediction methods are adopted to predict the mobile prediction data of the non-financial asset, and the final data has a certain deviation in a large probability, so that the mobile prediction data is consistent with the market prediction data of the prediction time period of the multiple market data prediction models in the embodiments of the present specification, the mobile prediction data of the non-financial asset can also be respectively predicted by different prediction methods in the embodiments of the present specification, and the preferred financial transaction strategy is adjusted based on the three predicted mobile prediction data of the non-financial asset, so that the adjusted preferred financial transaction strategy can meet the fund mobile stability under any one type of mobile prediction data of the non-financial asset.

If the prediction is performed by the residual fitting prediction method, only one residual fitting model is connected behind the basic prediction model (a model capable of predicting a specific value, such as a tree model), and the flexibility is high.

The foregoing describes specific embodiments of the present disclosure. Other embodiments are within the scope of the following claims. In some cases the actions or steps recited in the claims can be performed in a different order than in the embodiments and still achieve desirable results. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

Referring to fig. 5, fig. 5 is a schematic structural diagram of a transaction policy benefit prediction system according to an embodiment of the present disclosure.

As shown in fig. 5, the trading strategy revenue prediction system may include at least a market data prediction module 501, a revenue calculation module 502, and a model matching module 503. For convenience, model training module 504 is also shown in fig. 5. Wherein:

A market data prediction module 501, configured to predict and obtain a plurality of different return time period market prediction data in the same return time period using a plurality of market data prediction models;

the profit calculation module 502 is configured to obtain a real profit corresponding to the financial transaction policy according to the market history data of the return time period and the financial transaction policy; the system is also used for obtaining the return prediction benefits corresponding to the market data prediction models according to the market prediction data and financial transaction strategies of a plurality of different return time periods;

the model matching module 503 is configured to select a preferred market data prediction model adapted to the financial transaction policy according to each return predicted return and return real return;

the market data prediction module 501 is further configured to predict and obtain market prediction data of a future preset time period according to a preferred market data prediction model;

the profit calculation module 502 is further configured to obtain future predicted profits using the financial transaction policy in a future preset time period according to the future preset time period market prediction data and the financial transaction policy.

In one embodiment of the present description, the system further comprises:

the model training module 504 is configured to train the plurality of market data prediction models, and at least one of an internal architecture setting, a statistical feature extracted during training, a learning algorithm adopted during training, and training data adopted during training is different between any two market data prediction models in the plurality of market data prediction models.

In one embodiment of the present disclosure, the model training module 504 includes a feature extraction unit that is configured to extract raw feature parameters during the market data prediction model training process, and extract statistical features based on the raw feature parameters, where the statistical features include one or more of a maximum value, a minimum value, a mean value, and a standard deviation.

Referring to fig. 6, fig. 6 is a schematic structural diagram of another trading strategy profit prediction system according to the embodiment of the present disclosure.

As shown in fig. 6, the trading strategy revenue prediction system may include at least a market data prediction module 501, a revenue calculation module 502, a model matching module 503, and a financial trading strategy screening module 601. For convenience, model training module 504 is also shown in fig. 6. Wherein:

the profit calculation module 502 is further configured to obtain future predicted profits using the financial transaction policy in a future preset time period according to the future preset time period market prediction data and the financial transaction policy;

and the financial transaction policy screening module 601 is configured to select and obtain a preferred financial transaction policy used in a future preset time period according to future prediction benefits corresponding to all the financial transaction policies.

Referring to fig. 7, fig. 7 is a schematic structural diagram of another trading strategy profit prediction system according to an embodiment of the present disclosure.

As shown in fig. 7, the trading strategy revenue prediction system may at least include a market data prediction module 501, a revenue calculation module 502, a model matching module 503, a financial trading strategy screening module 601, an asset flow prediction module 701, a financial trading strategy cost prediction module 702, and a financial trading strategy adjustment module 703. For convenience, model training module 504 is also shown in fig. 7. Wherein:

the financial transaction policy screening module 601 is configured to select and obtain a preferred financial transaction policy used in a future preset time period according to future prediction benefits corresponding to all financial transaction policies;

The asset flow prediction module 701 is configured to predict and obtain non-financial asset flow prediction data in a future preset time period according to the non-financial asset flow history data;

the financial transaction policy cost prediction module 702 is configured to obtain a predicted cost of using the preferred financial transaction policy in a future preset time period according to the preferred financial transaction policy and the future preset time period market prediction data corresponding to the preferred financial transaction policy;

the financial transaction policy adjustment module 703 is configured to perform a proportional adjustment on the preferred financial transaction policy according to the asset balance, the non-financial asset movement prediction data, and the prediction cost of the preferred financial transaction policy.

In one embodiment of the present disclosure, the asset flow prediction module 701 predicts that the non-financial asset flow prediction data is a range of intervals.

In one embodiment of the present disclosure, the asset flow prediction module 701 predicts to obtain the non-financial asset flow prediction data by any one of a probability prediction method, a residual fitting prediction method, and a quantile regression prediction method.

It should be noted that, for each data of the input model and each data output by each module, a corresponding data quality detection report, for example, an anomaly detection report of the input data, is output through an additional quality detection module, and each report may be packaged and uploaded to a cloud storage module for storage, for example, OSS, for review.

The report of the abnormality detection of the input data may include a missing value, an abnormal value, a repeated value, and the like of the input data.

The quality check report of the model return test output data can comprise a plurality of different gap information between the return time period market prediction data and the return time period market history data, and a plurality of different gap information between the return test prediction benefits and the return test real benefits, namely the gap information of the prediction data and the real data output by the report.

Based on the content of the transaction policy profit prediction system described in the embodiments of the present disclosure, it can be known that the embodiments of the present disclosure provide a set of time sequence modeling flow frames for the transaction policy profit prediction and formulation process, and implement management capability for the whole user time sequence modeling flow, so that we can manage and optimize the flows from a unified perspective, including matching of prediction models in the process, profit prediction of transaction policies, and adjustment and optimization of the transaction policies. Thus, the workload of a large amount of adaptation codes which are required to be paid when using some time sequence prediction frames which are not provided with part of the content in the time sequence modeling flow frames provided by the embodiment of the specification is omitted. For example, when using a timing prediction framework such as Kats, glounTS, darts, a large amount of adaptation code effort is required to implement each flow of the scheme provided by the embodiments of the present disclosure.

In this specification, each embodiment is described in a progressive manner, and identical and similar parts of each embodiment are mutually referred to, and each embodiment mainly describes differences from other embodiments. In particular, for the trading strategy revenue prediction system embodiment, since it is substantially similar to the trading strategy revenue prediction method embodiment, the description is relatively simple, and reference is made to the description of the method embodiment in part.

Please refer to fig. 8, which illustrates a schematic structural diagram of an electronic device according to an embodiment of the present disclosure.

As shown in fig. 8, the electronic device 800 may include: at least one processor 801, at least one network interface 804, a user interface 803, memory 805, and at least one communication bus 802.

Wherein the communication bus 802 may be used to enable connectivity communication of the various components described above.

The user interface 803 may include keys, among other things, and the optional user interface may also include a standard wired interface, a wireless interface.

The network interface 804 may include, but is not limited to, a bluetooth module, an NFC module, a Wi-Fi module, and the like.

Wherein the processor 801 may include one or more processing cores. The processor 801 utilizes various interfaces and lines to connect various portions of the overall electronic device 800, perform various functions of the electronic device 800, and process data by executing or executing instructions, programs, code sets, or instruction sets stored in the memory 805, and invoking data stored in the memory 805. In the alternative, processor 801 may be implemented in at least one of the hardware forms DSP, FPGA, PLA. The processor 801 may integrate one or a combination of several of a CPU, GPU, modem, and the like. The CPU mainly processes an operating system, a user interface, an application program and the like; the GPU is used for rendering and drawing the content required to be displayed by the display screen; the modem is used to handle wireless communications. It will be appreciated that the modem may not be integrated into the processor 801 and may be implemented on a single chip.

The memory 805 may include RAM or ROM. Optionally, the memory 805 comprises a non-transitory computer readable medium. Memory 805 may be used to store instructions, programs, code, sets of codes, or instruction sets. The memory 805 may include a stored program area and a stored data area, wherein the stored program area may store instructions for implementing an operating system, instructions for at least one function (such as a touch function, a sound playing function, an image playing function, etc.), instructions for implementing the above-described respective method embodiments, etc.; the storage data area may store data or the like referred to in the above respective method embodiments. The memory 805 may also optionally be at least one storage device located remotely from the aforementioned processor 801. An operating system, network communication module, user interface module, and transaction policy benefit prediction application may be included in the memory 805 as one type of computer storage medium. The processor 801 may be used to invoke the transaction policy benefit prediction application stored in the memory 805 and to perform the transaction policy benefit prediction and formulation steps mentioned in the previous embodiments.

Embodiments of the present disclosure also provide a computer-readable storage medium having instructions stored therein, which when executed on a computer or processor, cause the computer or processor to perform the steps of one or more of the embodiments shown in fig. 2-4 described above. The above-described constituent modules of the electronic apparatus may be stored in the computer-readable storage medium if implemented in the form of software functional units and sold or used as independent products.

In the above embodiments, it may be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented in software, may be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When loaded and executed on a computer, produces a flow or function in accordance with embodiments of the present description, in whole or in part. The computer may be a general purpose computer, a special purpose computer, a computer network, or other programmable apparatus. The computer instructions may be stored in or transmitted across a computer-readable storage medium. The computer instructions may be transmitted from one website, computer, server, or data center to another website, computer, server, or data center by a wired (e.g., coaxial cable, fiber optic, digital subscriber line (Digital Subscriber Line, DSL)), or wireless (e.g., infrared, wireless, microwave, etc.). The computer readable storage medium may be any available medium that can be accessed by a computer or a data storage device such as a server, data center, etc. that contains an integration of one or more available media. The usable medium may be a magnetic medium (e.g., a floppy Disk, a hard Disk, a magnetic tape), an optical medium (e.g., a digital versatile Disk (Digital Versatile Disc, DVD)), or a semiconductor medium (e.g., a Solid State Disk (SSD)), or the like.

Those skilled in the art will appreciate that implementing all or part of the above-described embodiment methods may be accomplished by way of a computer program, which may be stored in a computer-readable storage medium, instructing relevant hardware, and which, when executed, may comprise the embodiment methods as described above. And the aforementioned storage medium includes: various media capable of storing program code, such as ROM, RAM, magnetic or optical disks. The technical features in the present examples and embodiments may be arbitrarily combined without conflict.

The above-described embodiments are merely preferred embodiments of the present disclosure, and do not limit the scope of the disclosure, and various modifications and improvements made by those skilled in the art to the technical solutions of the disclosure should fall within the protection scope defined by the claims of the disclosure without departing from the design spirit of the disclosure.

Claims

1. A trading strategy revenue prediction method, comprising:

2. The trading strategy return prediction method according to claim 1, further comprising the steps of, after predicting data of the market in a future preset time period and a financial trading strategy to obtain a future predicted return using the financial trading strategy in the future preset time period:

according to the same method, selecting a preferable market data prediction model which is adapted to other financial transaction strategies;

predicting and obtaining market prediction data of a preset time period in the future according to a preferable market data prediction model which is adapted to other financial transaction strategies;

market forecast data of a future preset time period and other financial transaction strategies are forecast according to a forecast model of the preferable market data respectively adapted to the other financial transaction strategies, so that future forecast benefits of using the other financial transaction strategies in the future preset time period are obtained;

And selecting and obtaining a preferred financial transaction strategy used in a future preset time period according to the future prediction benefits corresponding to all the financial transaction strategies.

3. The transaction policy profit prediction method according to claim 2, wherein after selecting the preferred financial transaction policy to be used in the future preset time period according to the future predicted profit of all financial transaction policies, the method further comprises:

predicting and obtaining non-financial asset liquidity prediction data in a future preset time period according to the non-financial asset liquidity history data;

according to the preferred financial transaction strategy and the market forecast data of the future preset time period corresponding to the preferred financial transaction strategy, forecast cost of using the preferred financial transaction strategy in the future preset time period is obtained;

and proportional adjustment is carried out on the preferred financial transaction strategy according to the asset balance, the non-financial asset liquidity prediction data and the prediction cost of the preferred financial transaction strategy.

4. A transaction policy revenue prediction method in accordance with claim 3, said non-financial asset flow prediction data being a span of intervals.

5. The transaction strategy profit prediction method according to claim 4, wherein the non-financial asset flow prediction data is predicted by any one of a probability prediction method, a residual fitting prediction method and a quantile regression prediction method.

6. The transaction strategy revenue prediction method of claim 1 or 2, wherein at least one of internal architecture settings, statistical features extracted during training, learning algorithms adopted during training, training data adopted during training are different between any two market data prediction models in the plurality of market data prediction models.

7. The method of claim 6, wherein the statistical features extracted during the training of the market data predictive model include one or more of maximum, minimum, mean, and standard deviation.

8. A trading strategy revenue prediction system, comprising:

9. The trading strategy revenue prediction system of claim 8, further comprising:

and the financial transaction strategy screening module is used for selecting and obtaining a preferred financial transaction strategy used in a future preset time period according to the future predicted gains corresponding to all the financial transaction strategies.

10. The trading strategy revenue prediction system of claim 9, further comprising:

the asset flow prediction module is used for predicting and obtaining non-financial asset flow prediction data in a future preset time period according to the non-financial asset flow history data;

the financial transaction strategy cost prediction module is used for obtaining the prediction cost of using the preferred financial transaction strategy in the future preset time period according to the preferred financial transaction strategy and the market prediction data of the future preset time period corresponding to the preferred financial transaction strategy;

And the financial transaction strategy adjustment module is used for carrying out proportional adjustment on the preferred financial transaction strategy according to the asset balance, the non-financial asset movement prediction data and the prediction cost of the preferred financial transaction strategy.

11. A trading strategy revenue prediction system in accordance with claim 10, the asset flow prediction module predicting that the resulting non-financial asset flow prediction data is a span of intervals.

12. The trading strategy revenue prediction system of claim 11, the asset flow prediction module predicts non-financial asset flow prediction data by any one of probabilistic prediction, residual fitting prediction, quantile regression prediction.

13. A trading strategy revenue prediction system according to claim 8 or 9, further comprising:

the model training module is used for training a plurality of market data prediction models, and at least one of internal architecture setting, statistical characteristics extracted in the training process, a learning algorithm adopted in the training process and training data adopted in the training exists between any two market data prediction models in the plurality of market data prediction models.

14. The trading strategy revenue prediction system of claim 13, the model training module including a feature extraction unit to extract raw feature parameters during the market data prediction model training process and to extract statistical features based on the raw feature parameters, the statistical features including one or more of maximum, minimum, mean, standard deviation.

15. An electronic device includes a processor and a memory;

the processor is connected with the memory;

the memory is used for storing executable program codes;

the processor runs a program corresponding to the executable program code by reading the executable program code stored in the memory for performing the method according to any one of claims 1 to 7.

16. A computer readable storage medium, on which a computer program is stored, which computer program, when being executed by a processor, implements the method according to any one of claims 1-7.